Dog–Dog Social Behavior: What Free-Ranging Populations Show

1. The Gap the Dominance Critique Left

1.1 What Was Dismantled

The applied claim that dogs seek rank over humans, and that human–dog problems are rank contests to be settled by physical confrontation, has no supporting evidence and produces interventions that make matters worse (which is the case against the construct as it is used in training). That conclusion is well supported.

What followed was less useful. Having removed the pack model, the field placed less emphasis on describing what dogs actually do with each other. Owners were told that dogs are not wolves and that dominance is a myth, and were left with little positive account of dog–dog behavior — which matters, because most owners have to manage exactly that: two dogs in a household, a meeting on a narrow path, a group at daycare.

1.2 What Was Not Dismantled

Two things survived the critique and are frequently lost in the retelling. First, dominance in ethology describes a property of a relationship, established by the direction of exchanged behaviors, not a character trait of an animal and not a motivation (Bradshaw, Blackwell & Casey, 2009). Second, when researchers went and measured behavior in free-ranging dog groups rather than arguing about the concept, they found structure — repeatedly, in more than one population.

Those findings do not resurrect the training doctrine. Nothing in them supports alpha rolls, eating first, or going through doorways ahead of the dog. They do mean the honest answer to "do dogs form hierarchies?" is not "no."

1.3 What This Article Covers

What the studied populations look like and why they differ; whether free-ranging dogs form groups at all; what was measured when hierarchies were reported and which measure held up; how age and leadership relate to rank; how reproduction and pup rearing are organized; what the wolf comparisons show about tolerance and cooperation; and what any of this transfers to a household with two dogs in it. One caveat runs throughout: the field literature is small, it is concentrated in a few research groups and a few populations, and almost none of it is experimental.

2. Which Dogs Were Watched

2.1 Four Different Populations

Most of the apparent disagreement in this area dissolves once the study populations are separated. Four are in play, and they are not interchangeable.

Suburban free-ranging dogs in Italy. Groups on the outskirts of Rome, observed over years, with known individuals and recorded dyadic interactions (Cafazzo et al., 2010; Bonanni et al., 2017). These are the populations in which pack structure has been documented most thoroughly.

Urban street dogs in India. Large populations in and around Kolkata and other cities, studied through censuses and field observation, living as scavengers in dense human environments (Sen Majumder et al., 2014; Paul & Bhadra, 2018).

Captive packs of dogs and wolves. Hand-raised, similarly kept animals at the Wolf Science Center in Austria, tested experimentally (Range, Ritter & Virányi, 2015; Marshall-Pescini et al., 2017).

Companion dogs. Neutered dogs in a daycare group (Trisko & Smuts, 2015), dogs in a research kennel group (van der Borg et al., 2015), and dogs cohabiting in households (Vékony & Pongrácz, 2024).

2.2 Why the Population Determines the Answer

These four differ in nearly everything relevant: whether resources are defensible, whether reproduction is possible, whether group membership is chosen or imposed, and whether humans control the food. A dog population with intact animals competing for mates and scavenging dispersed waste has different problems to solve from a group of neutered pets whose food arrives in bowls.

The temptation is to pick the population that supports the position one already holds. The more useful reading is that dog social organization is variable and responsive to conditions, which is itself the finding.

2.3 The Sample Problem

Worth stating early: a large share of what is known about free-ranging dog social structure comes from a small number of groups observed by a small number of teams. The Italian work rests on packs in one region; the Indian work covers far more animals but with less resolution on individual relationships. Independent replication across regions is thin, and the strongest claims in this article should be read with that in mind.

2.4 What "Free-Ranging" Actually Means

The term covers animals with very different relationships to people, and the differences matter for what their behavior can be used to argue. Village and street dogs live among humans, eat human refuse or handouts, and are usually unowned but not unfamiliar. Owned free-roaming dogs return to a household. Feral dogs subsist without direct human provisioning and are comparatively rare.

Almost all of the research discussed here concerns the first category. That is the right population to study, because it represents the majority of the world's dogs, and it is the wrong population from which to infer what dogs would do without humans. A pack in suburban Rome is not a wolf pack minus the wilderness; it is a social system built around a human-generated food supply.

3. Do Free-Ranging Dogs Live in Groups?

3.1 The Indian Answer: Mostly Alone, But Not Randomly

A population-level census study of free-ranging dogs in urban India modeled whether observed groupings could have arisen by chance. The dogs foraged largely singly, as efficient scavengers do, and formed groups that at first looked random. Closer inspection of who was with whom showed the associations were not random: adults associated with the opposite sex during the mating season, and juveniles stayed close to adults outside it (Sen Majumder et al., 2014).

The authors' summary is worth keeping: whether to be social is a matter of choice rather than chance in these populations. Sociality is present, facultative, and organized around reproduction and rearing rather than around foraging.

3.2 The Italian Answer: Stable Packs

The suburban Roman populations look different. Groups have stable membership over years, a recognizable structure, and coordinated collective behavior including group movement and conflicts with neighboring groups (Bonanni et al., 2017; Bonanni, Valsecchi & Natoli, 2010).

Both results are sound, and they describe different ecologies rather than contradicting each other. Where food is dispersed and comes in small units, scavenging alone is efficient and grouping buys little. Where resources are clumped and defensible, or where territory has to be held against neighbors, grouping pays.

3.3 The Human Variable

Free-ranging dogs do not live in an ecosystem that excludes people. Social network analysis of urban free-ranging dogs in India found that humans occupied a central position in the dogs' interaction networks — interactions with people were a larger part of the picture than interactions with other dogs (Bhattacharjee & Bhadra, 2020).

This has a consequence that is easy to miss. The most common social environment for the world's free-ranging dogs is one in which humans supply food, deliver both friendly and hostile treatment, and shape where dogs can be. Dog society, in much of the world, is partly shaped by human environments — which is the same conclusion the domestication literature arrives at from the other direction (where what domestication actually changed is more contested than usually presented).

3.4 Space, Neighbors and What the Groups Defend

Where packs exist, they occupy space and meet other packs, and those encounters are structured rather than chaotic. In the Roman population, conflicts between neighboring groups were regular events with recognizable participants and predictable escalation and withdrawal (Bonanni, Valsecchi & Natoli, 2010).

Much of the information exchange around territory happens through scent rather than through encounters at all, which is the cheapest way to communicate presence, sex and reproductive state without meeting (as the olfactory work sets out in detail). A great deal of dog–dog social behavior is asynchronous: one dog deposits information, another reads it hours later, and no interaction occurs.

This is worth stating because the pet-dog version of dog–dog social behavior is almost entirely face-to-face and almost entirely supervised, which is close to the least representative sample of the species' social repertoire that could be constructed.

4. Hierarchy: What Was Measured

4.1 Three Things Called Dominance

The word covers at least three distinct measurements, and most public argument conflates them.

Agonistic dominance is the direction of threat and aggression between two individuals. Formal dominance is the direction of ritualized status signals, chiefly submission, given regardless of context. Competitive ability is who actually gets the contested resource. A study can find one, two or none of these, and reporting only one of them as "a hierarchy" is where much of the confusion starts (a measurement problem that recurs throughout this field).

4.2 The Roman Findings

In the group studied by Cafazzo and colleagues (2010), the agonistic dominance hierarchy showed substantial linearity. Male dogs were prominently but not exclusively dominant within each age class, and rank correlated positively with age. Submissive–affiliative behavior met the criteria for formal submission signals, but not every dog produced them, so they could not order all individuals. Rank order predicted priority of access to food reasonably well, and relationships stayed stable across different competitive contexts.

The follow-up across five packs found linear hierarchies in every pack, rank positively related to age in all but one, and — the part that gets omitted — potentially injurious aggression was very rare, with agonistic interactions highly ritualized (Bonanni et al., 2017). Status in these packs sat mostly with older animals.

4.3 Submission Is the Better Measure

The most useful methodological result from this literature is that aggression is a poor status indicator and submission is a good one. In a group of 24 neutered companion dogs at a daycare facility, significant linear hierarchies emerged from highly unidirectional displays of submission, and submission was the most frequent, most unidirectional and most linear category of agonistic behavior — a better indicator of status than aggression or dominance displays (Trisko & Smuts, 2015).

A quantitative analysis in a group of domestic dogs reached a compatible conclusion: high posture and muzzle bite qualified best as formal dominance indicators, while body tail wag and lowered postures were the best formal submission indicators, with lowering of posture covering 91 percent of dyads (van der Borg et al., 2015).

The practical translation is uncomfortable for the usual reading. If you want to know how two dogs stand with each other, watch what the quieter one offers, not what the louder one does.

4.4 What a Hierarchy Is Not

It is not a motivation. Nothing in the field data shows dogs pursuing status as a goal; the structure is inferred from the direction of exchanged behaviors, and inferring a drive from a pattern is precisely the error the concept has been criticized for (Bradshaw, Blackwell & Casey, 2009).

It is not a trait either. A dog is not dominant; a dog stands in a particular relationship to another particular dog. The same animal can be consistently subordinate to one housemate and consistently dominant to the next, which is the same relationship-versus-trait distinction that keeps causing trouble elsewhere (as it does with aggression, which is also situational rather than a property of the dog).

And it is not a management plan. That a hierarchy can be measured in a group of dogs says nothing about whether a human should insert themselves into it, which is a separate claim with its own evidence — and that evidence points the other way (with confrontational methods carrying documented costs).

4.5 The Dispute That Is Still Open

Two positions have been argued in print, and both are more careful than their popular versions. Bradshaw, Blackwell and Casey (2009) hold that dominance is a property of relationships, that there is no evidence for a dominance trait in individual dogs, that in their own study of a freely interacting group no overall hierarchy could be detected, and that relationships are better explained by learned expectations about resources than by status. Schilder, Vinke and van der Borg (2014) hold that formal dominance is present in domestic dogs, expressed through context-independent unidirectional status signals, and that the construct therefore remains useful — while stating explicitly that a human enforcing dominant status carries considerable risk and should be avoided.

The disagreement is genuine and unresolved. What both sides agree on is the part that matters for practice: the applied doctrine built on the word has no support from either of them.

4.6 Why Aggression Is the Wrong Yardstick

There is a statistical reason for the finding that submission works and aggression does not. Serious agonistic events are rare, and rare events give sparse matrices: many pairs of dogs never produce one, so most relationships stay unscored and no ordering can be computed. Submissive signals are frequent, so they cover most dyads and their direction can be established.

There is also a conceptual reason. Aggression happens for many reasons that have nothing to do with status — pain, fear, a contested item, an unfamiliar approach — and a measure that lumps them together measures nothing in particular. Submissive signaling has a narrower function.

The consequence for anyone assessing two dogs is direct: counting fights tells you about the fights, and watching who defers tells you about the relationship.

5. Age Rather Than Force

5.1 The Age Gradient

Across the Italian packs, rank was positively related to age in nearly every group studied, and where it could be tested, age predicted dominance better than body size (Bonanni et al., 2017). The earlier group showed the same pattern in a sex- and age-graded form: males prominent within each age class, adults over subadults, subadults over juveniles (Cafazzo et al., 2010).

An age-graded structure is a different object from a fighting order. It is closer to the pattern in cooperatively breeding species, where group members follow experienced companions, than to a tournament in which the strongest animal takes the top position.

5.2 Conflict Is Mostly Ritual

The most consistent finding across these packs is how little damage occurs. Agonistic interactions were highly ritualized and potentially injurious aggression was very rare (Bonanni et al., 2017). Under conditions where dogs are unmanaged, uncastrated, competing for real resources and free to fight, they mostly do not.

That is the function a stable structure serves: it resolves the same questions repeatedly at low cost. When people describe dog groups as inherently conflict-prone, they are usually describing groups assembled by humans without stable membership, not free-ranging packs.

5.3 Where the Damage Does Happen

The exception is instructive. Serious fighting in dog groups is associated with unstable membership, uncertain relationships, and enforced proximity without escape options — conditions that are common in human-managed settings and rare in free-ranging packs.

Resource conflicts are the other case, and they follow the guarding pattern rather than a status pattern: what generalizes is the contingency that an approach predicts loss (which is where the guarding evidence sits). Reading a bowl conflict as a rank dispute misdirects the intervention.

5.4 Learning to Be Social With Dogs

Free-ranging puppies grow up surrounded by adult dogs of several ages, with continuous access to conspecific interaction and to correction that is delivered by animals that have no interest in escalating. Most pet dogs grow up with a fraction of that, in curated encounters with age-matched dogs, and with the human deciding when interaction starts and stops.

The developmental literature is clear that early social experience shapes later social behavior, and that the window in which it does so is limited (which is what the sensitive-period evidence supports). What it does not support is the confident timetable of later fear windows that circulates in training (where the developmental evidence is thinner than the popular account).

The honest position is that dog-directed social competence is built by experience with dogs, that the amount most pet dogs get is small compared with the free-ranging baseline, and that the precise scheduling claims made about it are not established.

6. Leadership Is Not the Same as Rank

6.1 Who Gets Followed

One of the more useful results in this literature separates two things routinely treated as one. In free-ranging dogs, leadership during collective movement was not the property of the top-ranking individual by agonistic measures. Dogs that received more formal submission in greeting contexts and had more affiliative relationships were more likely to take the group-leading role (Bonanni, Cafazzo, Valsecchi & Natoli, 2010).

Agonistic rank alone did not predict who the others followed. Two roles exist, and they can be held by different animals.

6.2 Costs Are Not Distributed by Rank Either

A companion study on conflicts between neighboring packs found that participation was individual: some dogs consistently took the risks of approaching rivals and others consistently hung back, and the ones who received more submission were not the ones who took more risk (Bonanni, Valsecchi & Natoli, 2010).

Free-riding within a group is a phenomenon that requires no cognitive story about fairness to be observable. It does mean that a high-status dog is not necessarily the group's defender, which is close to the opposite of the folk model.

6.3 Why This Matters for Practice

The applied version of the pack model rests on there being one position, held by one individual, that controls everything. The field data show at least three partly independent things: who wins contests, who is followed, and who takes risks. A dog can hold one and not the others.

Any household system built on identifying "the leader" and supporting them is therefore building on a construct the data do not describe (and it imports intentions the behavior does not license).

7. Reproduction and Rearing

7.1 Multiple Breeders, Non-Random Outcomes

Wolf packs are typically reproductively skewed, with breeding concentrated in one pair (Mech, 1999). Free-ranging dog packs are not: the Roman pack studied over several seasons comprised multiple breeding individuals. Even so, both male copulation success and female reproductive success were positively influenced by a combination of dominance rank, age and leadership (Cafazzo et al., 2014).

The structure is therefore neither wolf-like exclusivity nor a free-for-all. Rank does not decide who breeds, but it is one of several variables that shift the odds.

7.2 Who Raises the Puppies

A five-year field study of parental care in free-ranging dogs in India observed 23 mother–litter units across 15 groups and found widespread allo-parenting by both adult males and adult females (Paul & Bhadra, 2018). Care of young was not confined to the mother.

This is one of the clearer demonstrations that free-ranging dogs are more socially organized than the scavenger-loner picture suggests, and it sits awkwardly with the claim that domestication stripped dogs of cooperative behavior toward conspecifics.

7.3 What Follows for the Wolf Comparison

Wolves cooperate obligately: they hunt together, rear pups together and defend territory together, and failure to cooperate is costly. Dogs in most free-ranging populations scavenge individually and can rear young with variable amounts of help. The two species face different problems, and there is no reason to expect the same social solutions (which is also visible in what remains of the predatory sequence).

8. Tolerance and Cooperation: The Comparative Evidence

8.1 Wolves Cooperate Better Than Dogs

The most direct test compared similarly raised and kept wolves and dogs on a cooperative string-pulling task, where two animals must pull two rope ends simultaneously to obtain food. Wolves coordinated their actions and succeeded; dogs pulled in alternation and never solved it, despite comparable interest in the task. Cooperation in wolves was more successful between partners of similar rank and with a close social bond (Marshall-Pescini et al., 2017).

The popular domestication story predicts the opposite: that selection for tameness made dogs more tolerant and more cooperative than wolves. With conspecifics, and in this task, it did not.

8.2 Dogs Have the Steeper Hierarchy

The companion finding is equally counterintuitive. In pair-wise food competition tests with nine wolves and eight mongrel dogs raised and kept under identical conditions, dominant and subordinate wolves both monopolized food and both showed agonistic behavior; in dogs, those behaviors were the privilege of the high-ranking animal. Low-ranking dogs held back and accepted threats. Neither species showed much serious aggression (Range, Ritter & Virányi, 2015).

The authors read this as a steeper hierarchy in dogs than in wolves, and as evidence that wolves are tolerant enough for wolf–wolf cooperation, which may have been the basis on which dog–human cooperation was built.

8.3 What This Does and Does Not Establish

It establishes that the claim "domestication made dogs more tolerant and more cooperative with each other" is not supported by the comparative data available. It does not establish that dogs cannot cooperate: pet dogs manage joint activities routinely, and the researchers themselves note that owner training is a plausible reason.

The limitations deserve naming. Both studies used hand-raised captive animals in small numbers, wolves at the Wolf Science Center are not wild wolves, and a string-pulling apparatus is a long way from anything either species does naturally. This is a strong result in a narrow setting rather than a general statement about the species.

8.4 What Dogs Are Good At Instead

The comparative results are about conspecific cooperation, and they should not be read as a general deficit. Dogs attend to humans, learn from human demonstration, and form attachment relationships with people that are structurally similar to those measured in other species (as the attachment work shows).

They also acquire information from watching, including from other dogs, which is a form of social competence that a string-pulling task does not measure (within the limits the social learning evidence actually supports).

A defensible summary is that dogs are specialists in a different social problem from wolves. Wolves solve coordination with conspecifics; dogs solve living alongside humans and around other dogs without needing to coordinate with them. Neither is a deficit; they are different jobs.

9. What Transfers to Pet Dogs

9.1 Household and Daycare Groups Do Show Structure

The daycare group of 24 neutered dogs produced a linear hierarchy based on submission, and mixed-sex pairs were more likely to affiliate and less likely to show unidirectional submission and aggression than same-sex pairs (Trisko & Smuts, 2015). In cohabiting companion dogs, questionnaire-based rank scoring was validated against behavioral tests: items about formal and agonistic rank performed reasonably in a competitive test, while leadership-related items did not (Vékony & Pongrácz, 2024).

So structure exists in pet groups, owners can perceive parts of it, and the parts they perceive least reliably are exactly the ones the folk model emphasizes most.

9.2 What an Owner Can Actually Read

The measurable signal is directional submission: which dog consistently lowers posture, licks the other's mouth, passes under the head, wags with the whole body toward the other. That direction is stable and informative. Aggression is not: it is rare, contextual, and a poor index of anything durable (van der Borg et al., 2015; Trisko & Smuts, 2015).

Which means that the household dog most people would identify as dominant — the noisy one, the one that starts things — may well be the subordinate animal in the pair.

9.3 What the Evidence Supports Doing

Keep group membership stable where possible. Nearly everything good in the free-ranging data — ritualized conflict, low injury rates, predictable relationships — occurs in groups with stable membership over long periods. Rotating strangers through a space is the condition under which conflict rises.

Manage the resources rather than the ranks. Conflicts in pet groups cluster around contested items, doorways, resting places and human attention. Removing the contest resolves more than adjusting anyone's status ever will.

Watch for the same-sex pattern. Mixed-sex pairs affiliated more and showed less unidirectional agonistic behavior than same-sex pairs (Trisko & Smuts, 2015), and conflicts between cohabiting dogs are more commonly reported between dogs of the same sex. This is a probabilistic pattern rather than a rule, and it is worth knowing before adding a second dog.

Let age do its work. In free-ranging groups, older animals hold status and younger ones defer. Consistently supporting the younger dog against the older one, which is what owners often do out of sympathy, works against a structure the dogs would otherwise settle themselves.

And treat introductions as a training problem with its own transfer requirements, because two dogs getting along in the garden have not thereby learned to get along in the hallway (which is the ordinary generalization problem).

9.4 What the Evidence Does Not Support

None of the above licenses the training doctrine. There is no evidence that a human occupies a rank in a dog group, that eating first or preceding the dog through doorways affects anything, or that physical confrontation resolves status questions. Confrontational methods carry documented behavioral and welfare costs, and they generate exactly the fallout the free-ranging data do not show between dogs (which is the fallout side of the punishment literature).

Nor is there support for reading every dog–dog interaction through rank. Most interactions are not about status; they are about a resource, a threshold, a startle, or a dog that has learned that approaching means losing something.

9.5 Encounters on Lead

The most common dog–dog situation in a pet dog's life has no equivalent in any of the studied populations. Two unfamiliar dogs are brought into close proximity, both are physically restrained, neither can control distance, and the encounter is brief and unrepeatable.

Almost everything that keeps free-ranging encounters peaceful is missing there: the option to withdraw, the time to exchange signals at distance, and the stability of a known relationship. Lunging and barking in that situation are frequently read as dominance or as aggression, and are usually neither — the arousal and the restraint account for most of it (which is a different phenomenon from aggression).

The practical implication is that the free-ranging findings offer little direct guidance for lead encounters, beyond the general point that distance and the ability to disengage are what the dogs themselves use (and that arousal constrains what a dog can do).

10. Summary at a Glance

Free-ranging dogs are variably social, not uniformly asocial. Urban Indian populations forage largely alone but associate non-randomly around mating and rearing (Sen Majumder et al., 2014), while suburban Italian populations live in packs with stable membership (Bonanni et al., 2017).

Hierarchies have been measured, repeatedly. Linear agonistic hierarchies appeared in Roman free-ranging groups (Cafazzo et al., 2010), in all five packs of a later study (Bonanni et al., 2017), and in a group of neutered companion dogs (Trisko & Smuts, 2015).

Submission is the usable measure, not aggression. Unidirectional submissive signals were the most frequent and most linear indicator of status; lowering of posture covered 91 percent of dyads (van der Borg et al., 2015; Trisko & Smuts, 2015).

Rank tracks age, and conflict is ritualized. Rank correlated with age in nearly every pack, age predicted status better than body size where tested, and potentially injurious aggression was very rare (Bonanni et al., 2017).

Leadership is a separate role. Dogs that received formal submission in greetings and had more affiliative relationships led group movements; agonistic rank alone did not predict who was followed (Bonanni, Cafazzo, Valsecchi & Natoli, 2010).

Reproduction is shared but not equal. Packs contained multiple breeders, yet rank, age and leadership together predicted copulation and reproductive success (Cafazzo et al., 2014), and pup rearing involved allo-parenting by males and females (Paul & Bhadra, 2018).

Domestication did not make dogs outperform wolves in the tested conspecific cooperation tasks. Wolves coordinated and succeeded where dogs did not (Marshall-Pescini et al., 2017), and dogs showed the steeper hierarchy in food competition (Range, Ritter & Virányi, 2015).

11. Research Gaps and Critical Appraisal

The field literature rests on few populations and few teams. A large share of what is known about free-ranging pack structure comes from groups in one Italian region studied by one research group, and from Indian populations studied by one other. Regional replication is scarce, and the two bodies of work use different methods at different resolutions, which makes them hard to compare directly.

Almost none of it is experimental. Field observation establishes patterns and cannot establish causes. The claim that stable membership reduces injurious aggression, for example, is consistent with the data and has not been tested by manipulating membership.

The wolf comparisons use captive, hand-raised animals in small numbers. Nine wolves and eight dogs in one study (Range, Ritter & Virányi, 2015); comparable numbers in the cooperation work (Marshall-Pescini et al., 2017). These are careful studies with real limits, and hand-raised captive wolves are not a stand-in for wild ones.

The dominance dispute is unresolved in the literature, not settled in either direction. Bradshaw, Blackwell and Casey (2009) and Schilder, Vinke and van der Borg (2014) reach different conclusions about whether the construct is useful for companion dogs, and both positions are defended by researchers who agree on rejecting the applied doctrine.

Neutering status is an uncontrolled variable across the whole comparison. Free-ranging dogs are intact and compete for mates; most studied pet groups are neutered. Whether that changes the structure or only the stakes is not established, and it complicates every transfer from field to household (where the behavioral effects are less clear-cut than commonly assumed).

Breed composition is almost never controlled. Free-ranging populations are mostly mixed-ancestry village dogs; pet groups are frequently purebred and morphologically extreme, which affects both signaling and how signals are received (with breed explaining only a small share of behavioral variation).

No study has tested the household recommendations. The practical points in section 9 follow from the structure of the findings and from general learning principles. No controlled trial has compared, say, stable versus rotating group membership on conflict rates in pet dogs. That work does not exist.

12. Conclusion

The dominance critique was right about training and overreached about ethology. Dogs do form structured relationships with each other; those relationships are visible in the direction of submissive signals rather than in aggression; they are graded by age rather than settled by force; and they coexist with leadership and risk-taking roles that belong to different individuals. Where dogs live in stable groups, serious fighting is rare, which is the strongest practical argument against managing dogs as though conflict were their default state.

What none of this supports is the doctrine that was built on the word. A measurable hierarchy among dogs is not a licence for a human to enter it, and the evidence on confrontational handling points firmly the other way. The useful summary for practice is narrower and duller than either camp's slogan: keep groups stable, manage resources rather than status, read submission rather than noise, and treat each new context as untested.

Key Insights (Takeaways)

• The pack model was replaced by a vacuum, and the field data fill it differently than either camp expected. Free-ranging dogs in Rome live in packs with stable membership and linear hierarchies (Bonanni et al., 2017; Cafazzo et al., 2010); free-ranging dogs in urban India mostly forage alone but associate non-randomly (Sen Majumder et al., 2014). Dog society is variable and tracks conditions.

• Status shows in submission, not in aggression. Unidirectional submissive signals were the most frequent, most linear and most usable status indicator, with lowering of posture covering 91 percent of dyads (van der Borg et al., 2015; Trisko & Smuts, 2015). The loud dog in a household pair is often the subordinate one.

• Rank is graded by age and conflict is ritualized. Rank correlated with age in nearly every pack studied, age beat body size as a predictor where both were tested, and potentially injurious aggression was very rare among unmanaged, intact dogs competing for real resources (Bonanni et al., 2017).

• There is no single top position. Who wins contests, who is followed on group movements and who takes risks in intergroup conflicts are partly independent (Bonanni, Cafazzo, Valsecchi & Natoli, 2010; Bonanni, Valsecchi & Natoli, 2010). The folk model requires one role where the data show at least three.

• Domestication did not make dogs outperform wolves in the tested conspecific cooperation tasks. Wolves coordinated and succeeded on a cooperative string-pulling task while dogs pulled in alternation and failed (Marshall-Pescini et al., 2017), and dogs showed the steeper hierarchy in food competition (Range, Ritter & Virányi, 2015) — from captive, hand-raised animals in small numbers.

• For a household, the levers are membership and resources. Stable groups, managed contests, awareness of the same-sex pattern (Trisko & Smuts, 2015) and no human attempt to occupy a rank. None of the household recommendations has been tested in a controlled trial, and they should be held at that strength.

References

Bhattacharjee, D., & Bhadra, A. (2020). Humans dominate the social interaction networks of urban free-ranging dogs in India. Frontiers in Psychology, 11, 2153. https://doi.org/10.3389/fpsyg.2020.02153

Bonanni, R., Cafazzo, S., Abis, A., Barillari, E., Valsecchi, P., & Natoli, E. (2017). Age-graded dominance hierarchies and social tolerance in packs of free-ranging dogs. Behavioral Ecology, 28(4), 1004–1020. https://doi.org/10.1093/beheco/arx059

Bonanni, R., Cafazzo, S., Valsecchi, P., & Natoli, E. (2010). Effect of affiliative and agonistic relationships on leadership behaviour in free-ranging dogs. Animal Behaviour, 79(5), 981–991. https://doi.org/10.1016/j.anbehav.2010.02.021

Bonanni, R., Valsecchi, P., & Natoli, E. (2010). Pattern of individual participation and cheating in conflicts between groups of free-ranging dogs. Animal Behaviour, 79(4), 957–968. https://doi.org/10.1016/j.anbehav.2010.01.016

Bradshaw, J. W. S., Blackwell, E. J., & Casey, R. A. (2009). Dominance in domestic dogs — Useful construct or bad habit? Journal of Veterinary Behavior, 4(3), 135–144. https://doi.org/10.1016/j.jveb.2008.08.004

Cafazzo, S., Bonanni, R., Valsecchi, P., & Natoli, E. (2014). Social variables affecting mate preferences, copulation and reproductive outcome in a pack of free-ranging dogs. PLoS ONE, 9(6), e98594. https://doi.org/10.1371/journal.pone.0098594

Cafazzo, S., Valsecchi, P., Bonanni, R., & Natoli, E. (2010). Dominance in relation to age, sex, and competitive contexts in a group of free-ranging domestic dogs. Behavioral Ecology, 21(3), 443–455. https://doi.org/10.1093/beheco/arq001

Marshall-Pescini, S., Schwarz, J. F. L., Kostelnik, I., Virányi, Z., & Range, F. (2017). Importance of a species' socioecology: Wolves outperform dogs in a conspecific cooperation task. Proceedings of the National Academy of Sciences, 114(44), 11793–11798. https://doi.org/10.1073/pnas.1709027114

Mech, L. D. (1999). Alpha status, dominance, and division of labor in wolf packs. Canadian Journal of Zoology, 77(8), 1196–1203. https://doi.org/10.1139/z99-099

Paul, M., & Bhadra, A. (2018). The great Indian joint families of free-ranging dogs. PLoS ONE, 13(5), e0197328. https://doi.org/10.1371/journal.pone.0197328

Range, F., Ritter, C., & Virányi, Z. (2015). Testing the myth: Tolerant dogs and aggressive wolves. Proceedings of the Royal Society B, 282(1807), 20150220. https://doi.org/10.1098/rspb.2015.0220

Schilder, M. B. H., Vinke, C. M., & van der Borg, J. A. M. (2014). Dominance in domestic dogs revisited: Useful habit and useful construct? Journal of Veterinary Behavior, 9(4), 184–191. https://doi.org/10.1016/j.jveb.2014.04.005

Sen Majumder, S., Bhadra, A., Ghosh, A., Mitra, S., Bhattacharjee, D., Chatterjee, J., Nandi, A. K., & Bhadra, A. (2014). To be or not to be social: Foraging associations of free-ranging dogs in an urban ecosystem. Acta Ethologica, 17(1), 1–8. https://doi.org/10.1007/s10211-013-0158-0

Trisko, R. K., & Smuts, B. B. (2015). Dominance relationships in a group of domestic dogs (Canis lupus familiaris). Behaviour, 152(5), 677–704. https://doi.org/10.1163/1568539X-00003249

van der Borg, J. A. M., Schilder, M. B. H., Vinke, C. M., & de Vries, H. (2015). Dominance in domestic dogs: A quantitative analysis of its behavioural measures. PLoS ONE, 10(8), e0133978. https://doi.org/10.1371/journal.pone.0133978

Vékony, K., & Pongrácz, P. (2024). Many faces of dominance: The manifestation of cohabiting companion dogs' rank in competitive and non-competitive scenarios. Animal Cognition, 27(1), 12. https://doi.org/10.1007/s10071-024-01842-0